Alterego reported a scenario where the web server's cleanup() would wait

forever for 1 child thread (presumably the http_logging_thread) to terminate:

 >   8/27  02:39:43p  Waiting for HTTP logging thread to terminate...
 >   8/27  02:39:43p  HTTP logging thread started
 >   8/27  02:39:43p  HTTP logging thread received NULL linked list log entry
 >   8/27  02:40:44p  !TIMEOUT waiting for HTTP logging thread to terminate
 >   8/27  02:40:44p  Done waiting for HTTP logging thread to terminate
 >   8/27  02:40:44p  0000 Waiting for 1 child threads to terminate

 <infinite wait loop here>

- Add an extra terminate_http_logging_thread check to the top of the loop in
  http_logging_thread().
- Signal the log_list semaphore in the "Waiting for HTTP logging thread" and
  "Waiting for x child threads" loops.